Integration of optical components into a silicon-on-insulator (SOI) platform allows for the fabrication of a small size device, down to a submicron level, compatible with silicon electronic components and standard fabrication processes. Such optical components may have propagation modes of different dimensions (used interchangeably herein with “spot-sizes”) and shapes. For example, components such as waveguides, splitters and electro-optic modulators typically have a small spot-size of about 0.5 μm. Optical fibers have a larger spot-size of about 10 μm. These photonic components are coupled with each other via a spot-size converter to avoid energy loss due to their spot-size mismatch during the propagation from one optical component to another.
An existing spot-size converter that converts the spot-size of a first waveguide into the spot-size of a second waveguide simply involves a longitudinal transitional waveguiding structure between the two waveguides. When the waveguiding structure is large enough, the waveguiding structure confines light substantially all along the structure. The light confined in the first waveguide, having a spot-size similar to the size of the first waveguide, may gradually change the spot-size, as the light propagates through a transitional waveguiding structure, up to a size similar to the size of the second waveguide when the light reaches the second waveguide. The transitional waveguiding structure region may include changes in the width, height, or both. Such a simple waveguiding structure may be referred to as a tapered waveguiding structure. Typically, this type of a spot-size converter with a tapered waveguiding structure has a limited capability for converting a spot-size and is not sufficient to convert the spot-size from about 0.5 μm up to about 10 μm.
Another existing spot size converter involves a transitional waveguiding structure including a tapered region and a region where the waveguiding structure is not large enough to substantially confine the light. Similarly to the existing spot-size converter with a tapered waveguiding structure discussed above, the light confined in a first waveguide may pass through the transitional waveguiding structure where the width, height or both are decreased such that the waveguiding structure confines the light significantly less. Consequently, instead of decreasing, the spot-size rather increases and most of the light extends outside of the core of the waveguiding structure when the light reaches the second waveguide. This type of a waveguiding structure may be referred to as an inverted-taper waveguiding structure. Again here, this type of a spot-size converter with an inverted taper waveguiding structure has a limited capability for converting a spot-size and is not sufficient to convert the spot-size from about 0.5 μm up to about 10 μm.
Yet another type of a spot size converter may be composed of two such inverted taper waveguiding structures. Due to the poor confining capability at the tip of an inverted taper waveguiding structure, an interaction with another inverted taper waveguide in close lateral proximity, for example, placed side-by-side with their tapered tips in opposite directions, as illustrated in FIG. 1(a), may cause exchange of energy. Thus, the light from a first inverted taper waveguiding structure may be transferred in a second inverted taper waveguiding structure. A condition for an efficient exchange of energy between the two waveguiding structure is that their mode overlap is sufficiently large. Another condition for an efficient exchange of energy is that phase velocities (i.e. the effective refractive indices) in the two waveguiding structures substantially match. Due to the use of two different waveguides, this type of a spot-size converter with two inverted taper waveguiding structures may offer a better capability for converting a spot-size.
Fabrication of silicon optical devices involves successive deposition, treatment and partial etching of different materials on top of each other. In the etching process, chemicals may be used to preferentially etch one material while leaving another one virtually intact. Specifically, an etching process of the first material may be accurately controlled to stop when reaching the second material. The second material act as a so-called “etch stop” for the first material. In particular, a dielectric material such as silicon nitride can act as an etch stop for another dielectric material such as silicon oxide. As a result, thin layers of silicon nitride are often used as etch stop for silicon oxide within a stack of different layers containing different levels of metal circuits.